Pathos building repeatable acquisition pipeline
Pathos
This points to Pathos turning Foundry into a repeatable buying engine, not just a trial optimization tool. The key shift is that the platform is now screening outside assets, matching them to specific biomarker-defined patient groups, and moving them into owned programs where Pathos keeps more of the upside. DO-2 shows the model in practice, and management tied it to three other major portfolio decisions in early 2026, which suggests a steady sourcing process rather than a rare opportunistic deal.
-
DO-2 is a concrete example of the workflow. Foundry identified and evaluated DeuterOncology’s MET inhibitor, Pathos then bought a majority stake, and the asset came with Phase 1 work already completed across European sites. That means the platform is not just picking ideas, it is selecting near-clinic assets that can be pulled into Pathos control quickly.
-
This makes Pathos look less like a services style AI biotech and more like an asset aggregator in oncology. The economic difference is simple. Instead of earning limited value by helping someone else run a better trial, Pathos can own or control the molecule itself and capture downstream economics across development, combinations, and label expansion.
-
The pattern also fits how the rest of the pipeline is being built. Pocenbrodib is already being pushed through biomarker-led combinations in prostate cancer, while P-500 was licensed as a Phase 2 ready PRMT5 program with plans to use AI to find higher response subgroups and additional settings. The same data layer that guides development can also guide what to acquire next.
If this cadence continues, Pathos moves toward a portfolio model where Foundry continuously feeds new oncology programs into an owned pipeline. That would expand the company from improving clinical decisions on a handful of assets to systematically assembling a broader basket of mechanism specific cancer drugs, each with multiple indication paths and much larger economic ownership.